ALBERT

Description: Human prothrombinase assembled on synthetic membranes composed of phosphatidylcholine and phosphatidylserine (PCPS) catalyzes thrombin formation almost exclusively by sequential cleavage of prothrombin at Arg320 to yield the protease meizothrombin (mIIa) as an intermediate which is then further cleaved at Arg271. This cleavage pathway arises because Arg320 in intact prothrombin is cleaved ∼30-fold faster than Arg271. When prothrombin lacks γ-carboxyglutamic acid modifications (desGlaII), product formation is modestly decreased but bond cleavage largely occurs in the opposite order, yielding the zymogen, prethrombin 2 (P2), as an intermediate. This results from a reduction in the rate of cleavage at Arg320 in intact prothrombin and a partly compensating gain of function in cleavage at Arg271. Thus, membrane binding and/or other interactions mediated by γ-carboxyglutamic acids in the substrate play a major role in modulating the pathway for cleavage and whether the intermediate is a zymogen or protease. We now extend these approaches to evaluate prothrombinase function on activated platelets and human umbilical vein endothelial cells (HUVECs) at the physiologic concentration of prothrombin. Prothrombin activation was detected by measuring the appearance of product(s) with proteolytic activity and by analysis of the cleavage process by SDS-PAGE and western blot analysis quantitatively imaged by infrared fluorescence. This approach was validated by documenting equivalence in the progress curves obtained by western blot imaging or following staining of total protein in a purified system with PCPS membranes. Prothrombin cleavage was assessed using human platelets (108/ml), activated with thrombin and studied with saturating concentrations of Va and a limiting concentration of Xa. The pattern of prothrombin cleavage, intermediate accumulation and product formation was clearly distinct from that observed with PCPS membranes, indicating substantial flux towards thrombin formation via initial cleavage at Arg271 followed by cleavage at Arg320 producing the zymogen P2 as an intermediate. Transient formation of mIIa from initial cleavage at Arg320 was undetectable. Thus, prothrombinase assembled on thrombin activated platelets cleaves prothrombin in a way that is reminiscent of the cleavage of desGlaII rather than fully carboxylated prothrombin seen with PCPS membranes. In contrast, equivalent studies with prothrombinase assembled on thrombin activated HUVECs produced cleavage patterns consistent with significant flux towards thrombin formation via initial cleavage of prothrombin at Arg320 yielding mIIa as an intermediate. The use of desGlaII with HUVECs yielded lower rates and cleavage patterns similar to those obtained with fully carboxylated prothrombin in the platelet reactions. Our results document that the cleavage pathway for thrombin formation is dependent on cell type. Because mIIa is a protease with a different spectrum of activities from thrombin, its formation restricted to the vessel wall suggests an important regulatory role for the modulation of the pathway of prothrombin cleavage. Such bimodal regulation of the pathway for prothrombin cleavage on HUVECs and platelets suggests differential roles of prothrombinase assembled on platelets versus endothelium in regulating the hemostatic response to vascular injury.

Description: Prothrombin activation by prothrombinase is a paradigm for proteolytic activation reactions wherein product is produced following cleavage at more than one site in the substrate. Prothrombin is tethered to prothrombinase through exosite binding and is preferentially cleaved at R320. Subsequent cleavage at R271, in the resulting intermediate, yields thrombin. Selective presentation of the R320 site for active site docking and cleavage likely arises from the combined constraints of exosite-dependent tethering and the position of the cleavage site within the polypeptide sequence of prothrombin. A series of recombinant variants with substitutions in residues at and preceding the R320 cleavage site have been employed to investigate the role of geometric effects in enforcing the selective action of prothrombinase on this site in prothrombin. Replacement of the sequence D-G-R320 in wild type prothrombin (IIWT) with D-R-R320 (IIRR) or R-G-R320 (IIRGR) yielded substrate species that were indistinguishable from IIWT in their conversion to thrombin by prothrombinase. Equivalent progress curves for bands produced following initial cleavage at R320 followed by cleavage at R271 were established by SDS-PAGE for all three substrate species. Specific and quantitative cleavage at R320 rather than at R319 or R318 in the initial cleavage reaction was established by N-terminal sequencing. Substitution of R320 in IIWT with Q (IIQ320) rendered this site uncleavable and led to very slow cleavage only at the R271 site. These findings support a role for geometric constraints in precisely positioning R320 in II for preferential cleavage. Surprisingly, substitution of flanking sequences in IIQ320 to yield R-G-R-Q320 (II-1Shift) and R-G-R-G-Q320 (II-2Shift) yielded variants that were cleaved in a manner similar to IIWT. For either variant, the rate of prothrombin consumption was within 2-fold that observed with IIWT and could be quantitatively explained by cleavage at R319 (II-1Shift) or R318 (II-2Shift) determined by N-terminal sequencing. Thus, provided position 320 cannot be cleaved, geometric effects are not absolute and flexibility can be tolerated in the position of the scissile bond presented for preferential active site engagement. Variants containing RGR substitutions further shifted to the -3 and -4 positions were instead cleaved ~30-fold slower but at R271 similar to IIQ320. Dramatic loss of cleavage seen with these further N-terminal shifts establishes the limits of possible flexibility in substrate presentation to the active site. Further insights were provided by binding studies examining the ability of these variants to engage and displace 4-aminobenzamidine from the active site of the catalyst within prothrombinase assembled with XaA195. R320 in prothrombin engaged the active site with an equilibrium constant that was ~600-fold more favorable than R271. The register shift variants showed systematic decreases in their affinities for engaging the active site. Loss of cleavage at R316 in II-4Shift correlated with a markedly reduced affinity for active site docking. Thus, geometric constraints, arising from exosite binding, play an essential role in affecting the thermodynamics of active docking by prothrombin. However, very significant decreases in the equilibrium constant for active site docking of an otherwise highly favorable interaction are evident as only minor changes in rate. Large changes in rate and qualitative differences in cleavage pathway result when the equilibrium constant for cleavage at R316 in II-4Shift approaches that for R271. Our findings provide an explanation for how the action of prothrombinase on prothrombin is regulated by both precise substrate geometry and competition for active site docking between different sites within the substrate. The unexpected relationship between thermodynamics and rate reveals new mechanistic insights into the ability of prothrombinase to discriminate between sites within prothrombin and how thrombin formation is regulated.

Description: We have shown that human high molecular weight kininogen is proangiogenic due to release of bradykinin. We now determined the ability of a murine monoclonal antibody to the light chain of high molecular weight kininogen, C11C1, to inhibit tumor growth compared to isotype-matched murine IgG. Monoclonal antibody C11C1 efficiently blocks binding of high molecular weight kininogen to endothelial cells in a concentration-dependent manner. The antibody significantly inhibited growth of human colon carcinoma cells in a nude mouse xenograft assay and was accompanied by a significant reduction in the mean microvascular density compared to the IgG control group. We also showed that a hybridoma producing monoclonal antibody C11C1 injected intramuscularly exhibited markedly smaller tumor mass in a syngeneic host compared to a hybridoma producing a monoclonal antibody to the high molecular weight kininogen heavy chain or to an unrelated plasma protein. In addition, tumor inhibition by purified monoclonal antibody C11C1 was not due to direct antitumor effect because there was no decrease of tumor cell growth in vitro in contrast to the in vivo inhibition. Our results indicate that monoclonal antibody C11C1 inhibits angiogenesis and human tumor cell growth in vivo and has therapeutic potential for treatment of human cancer. (Blood. 2004;104:2065-2072)

Description: High molecular weight kininogen (HK) is known to bind specifically and saturably to Mac-1 with a Kd = 9–18 nM for neutrophils and to uPAR with a Kd =30 nM for endothelial cells. However, the functional results of HK interaction with Mac-1 or uPAR on leukocytes is not fully understood. Kallikrein cleavage of single chain HK to a two chain form (HKa) with release of bradykinin (BK) occurs in sepsis, arthritis, and inflammatory bowel disease. We hypothesized that HKa stimulates secretion of inflammatory cytokines. Mononuclear cells were isolated from normal subjects by a Histopaque density gradient. We have expressed kininogen domain 3 (D3) and a fragment of domain 3, coded for by exon 7, E7P (aaG235-Q292), in E. Coli as glutathione S-transferase (GST) fusion proteins. HK and HKa were purified proteins. GST was recombinant. All proteins contained